Re: Kernel Tracker / CeLog issues by henrik
henrik
Wed Nov 03 08:04:51 CST 2004
henrik.ramquist@transpondertech.se (Henrik R.) wrote in message news:<7dc4b0d.0410290437.69bad741@posting.google.com>...
> Yes, I've checked that it is implemented and supported (had problems
> with this first when I tried to run remote call profiler which also
> depends on this functionality).
>
> Is there some way of determine if CeLog is working properly, can it be
> problems due to something else not working as it should.
>
> /Henrik
>
> "bsqr_TSAT" <TSatagaj at hotmail dot com> wrote in message news:<uCLuuNGuEHA.2184@TK2MSFTNGP12.phx.gbl>...
> > Can you verify that OEMQueryPerformanceCounter() and
> > OEMQueryPerformanceFrequency() are implemented and supported on your device?
> > Your device should support some type of high resolution, 64-bit timer.
> > Without this support, you will see the errors that you're describing. The
> > information in Kernel Tracker will be accurate to the degree that the order
> > of thread calls will be shown correctly, but you won't be able to count on
> > the time stamps displayed.
> >
> > BSQR_TSAT
> >
> >
> >
> > "Henrik R." <henrik.ramquist@transpondertech.se> wrote in message
> > news:7dc4b0d.0410210148.3ff4ddf1@posting.google.com...
> > > Hi All,
> > >
> > > I have an issue which I've been looking for a solution for a couple of
> > > days (CE .NET 4.2).
> > >
> > > The problem is that Remote Kernel Tracker displays the following when
> > > I connect to our device (Kitl stream):
> > >
> > > "Remote Kernel Tracker detected invalid timestamps on some data while
> > > reading from the file or connection stream. Remote Kernel Tracker has
> > > corrected the timestamps to the best of its ability. Would you like to
> > > continue data collection?"
> > >
> > > It manages to get a few samples but these all seem to occur at the
> > > same time-instance.
> > >
> > > I've installed a patch from Microsoft which should fix a similar
> > > problem (Timestamp error in Remote Kernel Tracker KB826381), but it
> > > made no difference.
> > >
> > > There also seems to be a problem with the CeLog (I've tried to compile
> > > it and managed to get something out of it (.lib's) but it's seemed to
> > > make no difference either). I've tried to build the OS with different
> > > settings (such as "Flush Events To Release Directory", "Enable
> > > Profiling" etc.) but I get no output file called celog.clg in my
> > > release directory and I cannot connect using Remote Kernel Tracker.
> > >
> > > "CeLogFlush.exe" is present on the device and in the Registry/Init but
> > > I cannot se it in the active processes after I start.
> > >
> > > Could there be a problem in CeLog which results in no log/Remote
> > > Kernel Tracker not working? Anyone got any ideas on what it could be?
> > > What are the CeLog dependencies which could give these errors (Timers
> > > or...)? What can I do to test if CeLog is working properly?
> > >
> > > Best Regards,
> > > Henrik
Found the problem, It seems the company that modified the Xscale
package for us had removed the QueryPerformanceFrequency as well as
the QueryPerformanceCounter method which caused the program to output
"Invalid timestamps" see above.
/Henrik